California ISO. Discussion & Scoping Paper on Renewable Integration Phase 2. Renewable Integration: Market and Product Review Phase 2

Transcription

1 Discussion & Scoping Paper on Renewable Integration Phase 2 Phase 2 April 5, 2010

2 Issue Paper - Phase 2 Renewable Integration Table of Contents 1 Introduction Candidate Elements for the Comprehensive Market Design Roadmap Background Enhancements to Existing Market Design New spot market products Allocation of Integration Costs Modifications to Intra-Day Market Settlements Longer term procurement issues Comparison with other ISO/RTOs Pay for performance Regulation at NYISO, PJM and NEISO PJM Pay-For Performance Regulation Forward Capacity Markets Forward Reserve Markets Proposed Plan for Stakeholder Engagement Next Steps...25 CAISO/M&ID/M.Miller/L.Kristov Page 2 April 5, 2011

3 1 Introduction Over the past several years it has become clear that dramatic changes in the electricity sector will need to occur as a result of new public policies and regulations to reduce the environmental impacts of power production. Environmental policy is directly driving changes in the composition of the supply fleet mainly to increase the contribution of renewable resources and decrease reliance on fossil fuel resources which in turn presents new operational challenges, affects spot market outcomes (schedules, prices, revenues), and creates needs for new transmission infrastructure. Moreover, California s adoption of the 33 percent renewable portfolio standard ( 33% RPS ), requiring that 33 percent of the state s electricity consumption be supplied from renewable resources by 2020, allows a relatively short time frame for the industry to adapt to these changes, while the expected magnitude of the changes requires that the adaptations be designed carefully, from a whole-system perspective, and cost-effectively. Against this backdrop the ISO has already undertaken several initiatives to design, develop and implement needed changes in areas of its core responsibilities. Among other things, in July 2010 the ISO initiated the ( RI-MPR ), a stakeholder process to identify and develop potential changes to wholesale market design, including new market products and market rules to accommodate the expected substantial increase in production by variable energy resources (VER) over the next decade and address the expected revenue challenges that may face conventional resources needed to support integration of renewables. The RI-MPR was designed to be conducted in phases, so that nearterm operational needs could be addressed first, followed by market enhancements to address medium- and long-term needs. Phase 1 of the RI-MPR was begun in July and is now nearing completion; Phase 2 is hereby initiated with the present scoping and discussion paper. The ISO envisions two major objectives for Phase 2, and provides this paper to stimulate stakeholder discussion of these objectives and how best to achieve them. First, the ISO seeks to develop a comprehensive framework or roadmap for the market changes that will need to be designed and implemented over the next several years. Such a roadmap should take a longer-term view to identify the most effective and robust solutions to the new challenges and requirements rather than the most expedient or easy to implement, as well as a holistic view that takes into account the interactions between distinct market elements and rules. Such a roadmap should offer a vision of a well-functioning market end state, as well as a plan for getting there through a logically staged process for developing and implementing the component elements. The ISO believes that such a roadmap can and should be completed as a Phase 2 objective by the end of The second objective is to target specific market design changes or new market elements for completion, including approval by the ISO Board, by end of 2011 or early The ISO invites stakeholder input to help identify, from among the range of topics raised in this paper or others that stakeholders could propose for inclusion, a set of specific needs and issues that need to and logically can be addressed earlier rather than later, and that together would be feasible to complete in this time period. 1 Phase 1 is addressing a number of near-term incremental changes to market design to improve operational capabilities to support integration of renewables: reduction of the energy bid floor; refinements to the Participating Intermittent Resource Program (PIRP); limited modifications to bid cost recovery; and adoption of Regulation Energy Management (REM). The ISO s Revised Straw Proposal for Phase I is posted on the ISO website at CAISO/M&ID/M.Miller/L.Kristov Page 3 April 5, 2011

4 Following the stakeholder meeting to discuss the present paper and the ISO s receipt of written comments, the ISO will post a second issue paper that (1) provides a preliminary outline draft of the proposed comprehensive market design roadmap, laying out the intended scope of the roadmap without trying to fill in the substantive content that will be developed with stakeholders over the next several months, and (2) discussion of the high-priority topics identified for resolution by late 2010 or early 2011, including identification of the policy and design issues to be addressed and some of the options for consideration. The remainder of this paper provides: A survey of market design elements and issues that the ISO proposes to consider for inclusion in the first objective, the comprehensive roadmap described above, without any attempt to identify which of these should comprise the more focused design effort to be conducted for the second objective. A review of relevant market design elements that have been adopted by other ISOs and RTOs in the country. Proposed timetable for this initiative and some immediate next steps. 2 Candidate Elements for the Comprehensive Market Design Roadmap 2.1 Background The California ISO and other ISOs/RTOs will examine and implement a range of measures over the coming years to adapt the wholesale energy markets to the operational requirements of renewable integration. A number of significant market and operational changes could take place more or less concurrently, depending both on the quantity of renewable production as well as the particular mix of renewable resources and their locations. A number of trends appear to be inevitable. First, renewable energy will largely be displacing instate gas production in the California power system. The degree of that displacement will be a function of the mix of in-state and out-of-state renewable resources and the degree of dynamic scheduling of external resources. 2 The degree of displacement is also a function of the method and quantity of balancing capacity that will be necessary to support high level of renewable integration. The ISO s 20% RPS integration study showed that the displacement due to the combination of wind and solar causes the net load load minus wind and solar production to affect gas production across the day. 3 With a 33% RPS, on some days up to 50% of energy during the peak hours may be renewable (assuming high in-state development of solar resources). Hence, solar production may substantially displace gas-fired peaking units. Second, while renewable resources displace gas generator production, the gas fleet committed to provide energy and ancillary services will face an increasing number of starts and stops and will operate more often at minimum operating levels, compared to a benchmark scenario that does not include incremental renewable production. For example, the production simulations 2 3 Out of state resources that are firmed and shaped may displace other energy imports and thus have a less significant impact on in-state operations. See 20% RPS Study pg xiv CAISO/M&ID/M.Miller/L.Kristov Page 4 April 5, 2011

5 conducted for the 20% RPS study suggested that combined cycle plants would see a 35% increase in the number of starts compared to the benchmark scenario for That study also found that steam plants and gas turbines are expected to operate less often and have fewer starts. However, as system variability increases with as we move toward the 33% RPS, the ISO will likely need to procure additional reserves to provide load-following services. As this occurs, peaking conventional units may be needed more often to provide load-following service and incur an increased number of starts and stops. Third, wholesale energy prices will be affected in different ways by the new system conditions. First, significant renewable penetration will lead to a reduction in energy market prices in certain hours compared to what those prices would have been. Off-peak prices, for example, are expected to be reduced due to the higher wind production in those hours. How wind and solar production will affect on-peak prices will depend on the production efficiency of the gas plants being displaced from the peak hours. In addition, real-time prices overall will become more volatile, reflecting the variability in actual production by wind and solar in real-time. Key findings of the ISO 20% RPS Study indicate that in order to successfully integrate renewable resources the ISO will need increased operational flexibility. This will require additional ramp, load following and ancillary services capabilities from both generation and nongeneration resources. There will also likely be higher frequency and magnitude of overgeneration conditions. The ISO expects initial results of the 33% RPS study to be available during the second quarter of While this study will provide more details around the quantities of load following, regulation, ramping capacity, etc., needed to support the integration of variable energy resources; the 20% studies provide significant information on which to begin informed discussions with stakeholders on what types of changes to market design are needed. Given these concurrent market and operational impacts, the ISO seeks to consider as part of this initiative both 1) near term changes to existing market design that may provide the ISO with additional operational flexibility, as well as 2) longer term market design changes in the form of new spot market products and forward capacity products that will provide the ISO the needed operational characteristics from the resource fleet to integrate variable energy resources successfully. 2.2 Enhancements to Existing Market Design This section discusses potential changes to the existing market structure that could provide the ISO with additional operational flexibility to aid with the integration of renewable resources. These enhancements are currently listed in the Market Initiatives Catalog and were ranked medium in 2009 through the last round of the ranking process Hourly Contingency-Only Election for Operating Reserves Currently, due to software limitations, the contingency only designation for spinning reserve and non-spinning reserve must be for all 24 hours of each operating day. There is no ability to designate contingent or non-contingent reserves on an hour by hour basis. In addition, all incremental operating reserves procured in HASP or the RTPD are by default contingency only. This means that the ISO may dispatch these contingent reserves only when contingency conditions occur such as an unplanned outage, transmission contingency event or an imminent or actual system emergency. By providing hour by hour flexibility to the contingency-only CAISO/M&ID/M.Miller/L.Kristov Page 5 April 5, 2011

6 election the ISO could have more ancillary services capacity available to dispatch economically to address real-time imbalance energy requirements and other system conditions such as ramping constraints that will likely be more unpredictable with large quantities of intermittent resources under the 20% and 33% RPS Multi-Settlement System for Ancillary Services The existing market procures Ancillary Services in the day-ahead market to meet 100% of forecasted real time needs, and then procures additional A/S incrementally in real time only to the extent that they are needed due to changes in system conditions, loss of resources previously awarded AS, or demand exceeding the day ahead forecast. Currently, market participants do not have the ability to buy-back ancillary services they were awarded in the Day- Ahead market. Looking into the future where there will be more variability in real-time due to large volumes of renewable resources participating in the ISO markets, it may be beneficial to consider a multi-settlements system for ancillary services to provide more flexibility for the ISO and provide market participants with the ability to buy back and sell ancillary services closer to real-time Enhancements to RUC An aspect of RUC design that the ISO may want to examine as it prepares to accommodate a 33% RPS is how RUC will evaluate generation resources to ensure that there is enough capacity online, or available within a sufficiently short time frame, to avoid involuntary load shedding if the output of intermittent generation is significantly lower than the amount cleared in the day-ahead market. Hence, one element of the RUC design that needs to be evaluated is the methodology regarding wind and solar output assumptions that the ISO will use during RUC to assess whether enough capacity has been committed to meet the ISO s load forecast. The ISO s business practice manuals generally state that the ISO will make adjustments for differences between the amount of energy cleared in the IFM from intermittent resources and the ISO s forecast of their output, but the ISO may need to revisit how it determines these adjustments in light of larger volume of renewable resources coming online. As the level of intermittent resource output increases in future years, it may not be sufficient to simply adjust the RUC for the difference between the ISO s forecast of expected intermittent output and the amount of intermittent energy cleared in the day-ahead market. On the one hand, as the level of renewable output becomes larger, the magnitude of the potential forecast error (i.e., the potential difference between the actual and expected output) may become so large relative to the scheduled reserves that the ISO will need to account for the magnitude of this potential difference in the RUC commitment. At the same time, a conservative approach of running RUC based on an assumed low level of intermittent resource output in all hours of the day may greatly overstate the amount of energy required from energy limited resources, such as hydro, leading to the commitment of excess thermal generation in the RUC. A second element of the IFM/RUC commitment logic that the ISO and stakeholders might reevaluate is how the ISO should best ensure that it has enough ramp capability committed to accommodate the intra-hour variability of renewable generation output. This is conceptually distinct from the capacity question above, because it involves commitment of enough rampable generation to handle the potential level of variation in output within an hour, including downward ramp. One of the topics that should be considered is whether this evaluation is best structured within the IFM, within the RUC process, or potentially in a new market process. CAISO/M&ID/M.Miller/L.Kristov Page 6 April 5, 2011

7 Another issue to consider is whether or not performing RUC and IFM simultaneously would provide benefits as the ISO moves towards higher volumes of renewable resources. Currently, the ISO manages RUC in a sequential run that clears against the ISO load forecast after the IFM has concluded. Simultaneous RUC and IFM would reflect the ISO s requirements to have sufficient capacity and import energy available to operate the system in real-time as part of the day-ahead prices. This approach may also provide the ISO with more tools to address overgeneration conditions. For example, the current RUC design is not able to de-commit units when too much capacity was committed in the IFM run relative to the load forecast, whereas a simultaneous IFM and RUC would avoid this problem. While this more efficient unit commitment may provide benefits, it would not be a simple change to implement; the ISO believes therefore that more discussion is needed to determine if this enhancement provides enough benefits to be considered a higher priority over other enhancements needed to support renewable integration. 2.3 New spot market products Pay for Performance Regulation The 20% RPS study identified a need for additional regulation capacity particularly in the evening and morning ramp hours. For example in the summer season, the total simulated requirement of regulation up in 2012 (the total MW of the values plotted in the frequency distribution for 2012) as compared to 2006 was 37% greater. For regulation down the simulated requirements showed to be 11% greater than In the stakeholder initiative on Regulation Energy Management (REM), the ISO committed to explore further enhancements to regulation products as part of this Phase II initiative. Resources that more accurately follow AGC signals can reduce the amount of regulation the ISO must procure to meet system needs which could provide long-term cost benefits by eliminating or reducing the need for the ISO to increase ancillary service procurement targets due to intermittent resource penetration. Therefore, the ISO is considering a pay for performance product. The ISO also evaluated what the other ISOs are doing in regards to pay for performance regulation which is described in Section 3 of this document. In addition, FERC also issued a NOPR which is explained in the following paragraphs that proposes to change the rules for how ISOs compensate resources providing regulation. On February 17, 2011, FERC issued a notice of proposed rulemaking (NOPR) to address undue discrimination in the procurement of frequency regulation service (regulation) in organized wholesale electricity markets. The rules proposed by FERC would change how the ISO compensates resources providing regulation. 5 FERC asserts that faster ramping resources provide more Area Control Error (ACE) correction to system operators than slower ramping resources but do not appear to receive compensation 4 5 See 20% RPS study, pg 52, table 3.3 at: Regulation is used to control the energy output of generating units within a prescribed range in response to changes in system frequency, tie-line loading, or the relation of these to each other so as to maintain the target system frequency and/or the established interchange with other balancing authority areas within predetermined limits. The ISO s AGC system issues instructions every four seconds to each resource providing Regulation to help assure that CAISO balances demand and supply between five minute real-time dispatch intervals in the ISO Balancing Authority Area in accordance with NERC and WECC operating criteria. CAISO/M&ID/M.Miller/L.Kristov Page 7 April 5, 2011

8 for the service they provide. FERC states that certain ISOs and RTOs net regulation up and regulation down energy payments provided by resources and this compensation method does not acknowledge the greater amount of ACE correction provided by faster ramping resources. FERC is concerned that slower ramping resources may receive the same compensation as faster ramping resources under a netting approach even though the faster ramping resource provides more ACE correction to a system operator during a market interval. FERC also states that some ISOs and RTOs dispatch faster ramping resources earlier than slower ramping resources yet pay all resources the same rate for capacity and the same price for MWh of net energy. FERC s concern indicates that such compensation approaches may be unduly discriminatory to faster ramping resources. The ISO seeks comments from stakeholders on issues raised in FERC s NOPR as part of phase 2 of its renewable integration market and product review. FERC Compensation Proposal FERC proposes to require all resource bids to include opportunity costs and that all cleared bids for regulation receive a single market clearing price that reflects the total marginal costs of the marginal cleared resource. FERC proposes that ISOs and RTOs change their tariffs so that regulation resources receive a two-part payment. Two-part Payment FERC proposes a two-part payment structure for regulation that would include a capacity payment and a performance payment with an accuracy adjustment. Capacity Payment FERC proposes that regulating resources receive a capacity payment to hold capacity in reserve and not participate in the energy market. The payment must include the marginal regulating resource s opportunity costs and the ISO or RTO must calculate cross-product opportunity cost resulting from not participating in the energy market. 6 Where appropriate, FERC also states that resources should be permitted to include inter-temporal opportunity costs in their capacity bid (i.e. the foregone value if a resource must operate at one time and forego a profit from selling energy at a later time or incurring costs to consume energy at a later time). 7 Under the ISO market design, the ISO co-optimizes energy dispatch and reserve procurements in both the day-ahead market and the real-time market allowing a regulating resource within an ancillary services region to earn the marginal resource s opportunity cost, which also reflects cross-product opportunity costs. The ISO currently co-optimizes energy and ancillary services over multiple intervals. The ISO solicits comments on whether its capacity payments for regulating resources should also include inter-temporal opportunity costs. For storage devices, the inter-temporal opportunity cost represents the value that the resource forfeits to provide regulation up or regulation down 6 7 FERC Order 745 at pg. 23, FN 49 FERC indicates that the cross-product opportunity cost is the revenue a regulation provider loses because it is on stand-by to provide regulation and is not providing energy. FERC Order 745 at pg. 23, FN 50 explains that the inter-temporal opportunity cost is the trade-off presented to a thermal storage operator who would prefer to charge when prices are low. If such a resource were to provide frequency regulation, it could be asked to stop charging during low price periods and then be forced to charge during high price periods. CAISO/M&ID/M.Miller/L.Kristov Page 8 April 5, 2011

9 because it will have less flexibility to use its capacity for energy charging and discharging and take advantage of energy price differences between intervals. To reflect inter-temporal opportunity costs, the ISO would need to build additional constraints into the ISO optimization algorithm. The ISO seeks specific proposals on how it would account for inter-temporal opportunity costs in capacity payments for regulation for both storage resources and conventional resources. Issues to Consider- Capacity Payment Is there a minimum required amount of stored energy for a given interval of time for a storage device to qualify to provide regulation service? In real-time, should resources awarded to provide regulation in subsequent intervals be disqualified from providing regulation in subsequent intervals if the resource s stored energy falls below a minimum energy threshold due to energy releases in previous intervals? How would the ISO account for inter-temporal opportunity costs in the price of regulation energy for a storage device given the price could be different than the price of regulation energy provided by a conventional resource due to potential intertemporal constraints applied only to storage? Performance Payment with Accuracy Adjustment FERC proposes that regulating resources receive a performance payment based on the amount of up and down movement, in megawatts, a resource provides in response to the ISO/RTO AGC signal. FERC proposes the payment should also incorporate a resource s accuracy in following the AGC signal to provide ACE correction. A performance payment with an accuracy adjustment allows for resources that respond accurately to an AGC signal to receive greater compensation than resources that do not respond accurately. Historically, the ISO has qualified and compensated regulation resources for their ability to respond to an AGC signal; however, FERC is suggesting that the quality of regulation service provided is worthy of compensation beyond simply a resource s capability to provide regulation. 8 Mileage payment In addition to a capacity payment, FERC proposes that ISOs/RTOs compensate regulating resources with a mileage payment. A mileage payment would provide greater compensation to a resource that is able to perform more work, i.e. can travel a greater distance in a set amount of time relative to another resource. Under a mileage payment, FERC proposes that the ISO would sum the total absolute value of a resource s up and down movement multiplied by the price per MW of ACE correction. FERC proposes that the mileage payment reflect bids in the form of a price-per-mw of ACE correction ($/MW-ACE Correction). Under FERC s proposed rule, the ISO would determine the least cost set of resources based on available capacity, ramp rate, and price-per-mw ACE correction, thereby establishing the final ancillary service marginal price for regulation on the marginal regulating resource. The price-per-mw of ACE correction would be a separate and 8 The ISO s Regulation non-compliance program only rescinds Regulation capacity payments when the quantity of Regulation capacity is non-compliant. The ISO Regulation non-compliance program does not measure the accuracy of a resource responding to the AGC signal. (CAISO BMP for Compliance Monitoring, pg. 36.) CAISO/M&ID/M.Miller/L.Kristov Page 9 April 5, 2011

10 distinct bid from a resource s regulation capacity bid. In the alternative, if a market solution is not workable, FERC suggests the mileage rate could reflect an administrative rate. Accuracy adjustment FERC proposes that ISOs and RTOs measure the accuracy of a resource using currently available telemetry and that a resource only receive payment for regulation energy that actually corrects ACE. FERC states that there is little consensus concerning how to incorporate regulation accuracy into wholesale electricity market designs. The ISO solicits comments from stakeholders concerning developing mileage payment for regulating resources that includes an accuracy adjustment. Specifically, the ISO requests in put on the following questions: Issues to Consider- Performance Payment with an Accuracy Adjustment Does the fact that the ISO procures regulation up and regulation down as separate services have an impact on how the ISO would implement a performance payment? Are there minimum threshold performance standards to be eligible to receive a performance payment? Is there a correlation between fast ramping and accuracy? For instance, can a single fast ramping regulating resource be more accurate in satisfying ACE correction than several slower ramping regulating resources? How would stakeholders define accuracy? Should physical constraints of certain resource types be considered in an accuracy adjustment? For example: o For a hydro generator, after sending out a raise MW command, the MW output will lower before moving up to the target level. This is a characteristic of a hydro generator. o A fast ramping rate does not always mean fast response over a short time period even though there may be a correlation. Time constants of certain resource types, e.g. turbines, play an important role in response times over short periods. Do such time constant constraints impact the ability of turbines to accurately follow an AGC signal? How might the ISO apply an accuracy adjustment? Net Energy FERC questions whether net energy should be maintained in ISO and RTO regulation markets in light of its proposed two-part payment compensation method described above. The ISO calculates the regulation energy provided from the two 5-minute dispatch intervals over a 10-minute settlement interval. The ISO solicits comment on whether establishing a mileage payment would make the netting of energy across the settlement interval a moot point. Conclusion The ISO looks forward to engaging stakeholders in policy and technical discussions concerning the potential development of compensation approaches for regulating resources. Ultimately, the ISO wants to establish policies and market based incentives that will enable and encourage resources to more accurately follow the ISO AGC signal. Monetizing performance and accuracy to address ACE correction could assist the ISO with integrating greater amounts of variable CAISO/M&ID/M.Miller/L.Kristov Page 10 April 5, 2011

11 energy resources and could encourage new types of regulating resources. However, all benefits must be weighed against their cost, and so the ISO will seek input from stakeholders on costs and benefits of compensation methods and any related cost allocation concerns Load Following Reserve A key finding of the 20% RPS report was that load following requirements will increase substantially in some hours. A load-following requirement is defined as unloaded upward dispatch capability and unconstrained downward capability in the units committed through the day-ahead market and real-time market procedures to meet the schedule deviations in real-time by load and variable energy resources. As noted in the 20% RPS report, the ISO currently has no explicit load following requirement constraint but rather balance system needs through unconstrained dispatch capability available through current market procedures. Without an explicit load following constraint, if imbalance conditions change due to load and supply deviations more than the available capability is able to follow, imbalance shortages will arise and the ISO will have to lean on regulation resources or others in the interconnection to balance system needs. At the same time, some of the changes currently in progress such as the flexible ramp constraint will effectively better ensure sufficient load-following capacity. The 33% RPS study will provide more information as to additional quantities needed by 2020 to manage the grid. At some point, under criteria to be considered in this initiative, an explicit capacity reservation may be needed. Once the exact needs are better defined a number of design issues would need to be addressed. Requirements Since there is no WECC MORC requirement for a load-following reserve the CAISO and stakeholders would have to determine how and when to set requirements for this new product. Procurement The load following reserve could be procured regionally, sub regionally or both. The methodology for procurement would also need to be determined. The following options could be considered: 1) Incorporate procurement with other Ancillary Services in the Day-Ahead clearing process. This option would allow the load-following reserve to be co-optimized along with energy and other ancillary services bids. 2) Purchase only in HASP after Day-Ahead outcome is known. This approach would allow the CAISO to determine requirements for load-following reserves after the outcome of the IFM and RUC processes. This could ensure a more accurate determination of the requirements closer to Real-Time. 3) Procure through monthly or annual auction. It may be more efficient to procure load following reserve on a more forward, longer-term basis, to provide greater day to day certainty to the ISO about the availability of these reserves and to the qualified resources regarding their revenues from capacity payments. Further discussion of this concept is provided later in this paper in connection with a potential forward reserve market. CAISO/M&ID/M.Miller/L.Kristov Page 11 April 5, 2011

12 2.3.3 System Inertia and Frequency Response As larger volumes of variable energy resources displace more and more synchronous generation, there will be times when the synchronized inertia on the system could have a negative impact on the established stability limits of the system. During high variable energy production levels and low system load (off-peak hours or during weekends or holidays), there may not be enough rotating mass or inertia on the system to arrest frequency decline and/or enough governor response to stabilize the system frequency following a contingency. This could result in under-frequency relays picking up and disconnecting load from the system. How to potentially procure and price this operational need will require further discussion. The ISO is currently working on an intertia study with GE that will provide more insight into the magnitude of this potential concern. The inertia study is planned to be completed at the end of July Flexible Ramping Constraint The ISO has identified a near term operational need to ensure there is enough MW ramping capacity up and down to handle the range of net load 9 variations that can occur within the realtime operating hour. Currently the Integrated Forward Market (IFM) and the Real-Time Pre- Dispatch (RTPD) optimize resources based on a single point-forecast imbalance amount for each interval of the designated time horizon; i.e., each hour of the next day for the IFM and each 15-minute interval for the RTPD time horizon. 10 The optimization assumes both that there is no error in the net load forecast and that system conditions are constant within each interval. There are times when IFM and RTPD optimize resources so efficiently that they leave little online available, unscheduled capacity for use by the 5-minute Real-Time Dispatch (RTD) to meet any variation between the constant forecast conditions assumed in IFM and RTPD and what is actually occurring in real time. Given the nature of these assumptions used in the RTPD, real-time deviations from the net load forecast frequently require the RTD to dispatch capacity that can ramp to the needed target operating level by the next 5-minute interval in order to maintain reliable operation, particularly during hours of high load ramping. When these needs occur and the available supply of ramping capacity is not sufficient, the only tools left for the operator to manage the problem are to bias the forecast and/or issue exceptional dispatch instructions. To provide ISO operators the ability to manage these situations through the RTD market optimization, an additional constraint will be introduced in the IFM, RTPD and RTD optimizations, referred to as the upward and downward System Ramping Constraint (SRC). These new constraints will ensure that the optimization solutions in IFM, RTPD, and RTD make available some un-loaded upward ramping capacity below Pmax and some loaded downward ramping capacity above Pmin on the scheduled, internal, dispatchable, ramp-capable resources. These constraints will effectively ensure that there is sufficient capacity that is available to respond to real-time dispatch instructions and is capable of ramping to manage the variability caused by the changing system conditions and various numerical and forecasting model errors between the IFM, the RTPD and the corresponding RTD intervals that will occur starting roughly 15 minutes after each RTPD The net load forecast is the forecast of ISO internal load net of the forecast VER energy supply. RTPD looks ahead between four and seven 15-minute intervals to ensure there is sufficient capacity to meet demand over this multi-interval trajectory. CAISO/M&ID/M.Miller/L.Kristov Page 12 April 5, 2011

13 Because of the near-term operational need for the real-time ramping capability these new constraints will provide, the development of the design details of these constraints is not a topic for the initiative discussed in this paper. This initiative will, however, consider how the needs driving the implementation of the flexible ramping constraints could be met through a more market-based approach. The ISO expects that the intra-hour variability that these new constraints are designed to address will increase in the future as more VER are integrated into the ISO system. Therefore the ISO will implement the new constraints as a near-term, interim solution, until the ISO and stakeholders can develop a long-term approach through the present initiative that will appropriately value, procure and compensate the needed ramping service through the market structure Reflecting Constraints in Market Prices In addition to the physical constraints of the transmission system and generation resources, the ISO market optimizations may need to incorporate additional constraints to ensure that reliability requirements are met through market outcomes. The previously implemented Minimum Online Commitment (MOC) constraint and the pending Flexible Ramping Constraint are examples of constraints used to ensure that the market outcome will meet reliability requirements. In both constraints, the issue is that absent the constraint the ISO is not assured that the market outcome will provide sufficient resources to manage identified reliability concerns. In order to assess the appropriate pricing impact of the constraint, one must determine why the market did not commit the resources necessary to meet the operational requirement absent the constraint binding. Next one must evaluate the economic impact on the resource that has satisfied the constraint and on the other resources participating in the market. Then one must determine which components of the LMP should be impacted when the constraint is binding. Finally, one must determine if and how to price the constraint through either a direct capacity payment (spot market product) or calculation of opportunity costs. There may also be intermediate short-term steps such as paying a resource the shadow price of the constraint while the impact of the constraint is analyzed during actual market operations or a new product is being designed and implemented. Another approach could be to incorporate certain types of constraints into the annual local capacity requirements for Resource Adequacy (RA), so that the needed resources will be procured to meet RA requirements and receive RA capacity payments. The MOC constraint addresses the operational needs of certain operating procedures that require a minimum quantity of committed online resources located within a defined area in order to maintain reliability. Prior to implementing this constraint, the specific resources were committed outside the market through Exceptional Dispatch. Additional documentation is available in the Technical Bulletin available at Currently the shadow prices of the MOC are not incorporated directly into any pricing calculations, but if the constraint is regularly binding at high capacity levels it may warrant designing a mechanism to ensure the constraint impacts pricing. Absent this constraint, the specific resources in the defined area are not committed as other lower priced resources outside the defined area will meet system needs while satisfying the objective to minimize overall system costs. When the constraint is enforced, the needed resource inside the defined area is committed at least at its minimum load (PMin). However, if the resource is economic to provide energy above its minimum load the resource will be dispatched above PMin and be eligible to set the price. If however, the resource is not economic above based on its energy above minimum load the resource will not set prices because its energy bid is not cleared in the CAISO/M&ID/M.Miller/L.Kristov Page 13 April 5, 2011

14 market. If the resource has a revenue shortfall from operating at PMin it is eligible for bid cost recovery. The impact for some resources outside the defined area is they are not being committed. In addition, the commitment of the needed resource at PMin provides additional price taker supply, shifting the supply curve and potentially resulting in a lower or higher system marginal cost of energy (SMEC) relative to the SMEC without enforcement of the MOC constraint. In many cases the MOC constraint addresses constraints that are already reflected in the local capacity requirement in the annual Resource Adequacy requirements for the defined area, so that the specific resources committed by the MOC constraint already receive RA capacity payments. However, to the extent an MOC constraint is requiring non-ra resource to be committed, one may question if such non-ra resources have sufficient opportunity to be compensated for their service. The MOC does not lend itself to a spot market product because only specific resources are able to resolve the constraint; however, potentially an opportunity cost could be calculated that reflects the commitment of higher cost resources in the SMEC for all resources. The Flexible Ramping Constraint is being implemented because the ISO has observed that in certain situations reserves and regulation service procured in the real-time (RTPD and HASP) and units committed for energy in the fifteen minute unit commitment process (RTPD) lack sufficient ramping capability and flexibility to meet conditions in the five minute market interval. The issue can be addressed in part by enforcing the Flexible Ramping Constraint to ensure that sufficient upward and downward ramping capability of dispatchable resources is committed to enable the real-time dispatch (RTD) to follow load efficiently and reliably over an estimated range of potential variability of net load around the forecast. Additional documentation is available in the Technical Bulletin available at The Flexible Ramping Constraint does not offset or reduce the required procurement of regulation or contingency reserves and is not reflected in the energy price unless imbalance condition change such that in RTD the resource resolving the constraint is dispatched above minimum load. Since energy prices during RTPD and HASP are advisory for internal resources, the commitment of capacity in these real-time market processes does not directly impact realtime energy price. Incremental ancillary services are procured in RTPD and these prices are binding. In assessing the pricing impact of this constraint it is necessary to separate based upon the ramping direction. In instances where the Flexible Ramping Up Constraint is binding, an additional resource (presumably higher commitment plus bid cost) is committed at PMin and the resource which would have been committed and loaded up to its PMax absent the constraint is committed at a lower level so that the resource may be able to provide additional upward dispatch capability. The resources committed via the constraint are eligible to set prices in realtime if needed. The resources will be able to receive additional market revenue if the imbalance energy requirements increase above forecasted levels such that the need for ramping requirements are realized. Both resources are eligible for bid cost recovery to address any shortfalls from being committed. This is analogous to procuring additional non-contingent spinning reserves; however, the reserved capacity is used to meet real-time upward dispatch requirements and not to satisfy WECC contingency reserve procurement targets. If additional spinning reserves would have been procured in the day-ahead market (above WECC procurement targets) the additional capacity would have caused the spinning reserve Ancillary Service Marginal Price (ASMP) either to remain the same or increase and the reserve demand would be increased to reflect the reserved capacity. Due to the co-optimization of energy and ancillary services, there may be interplay with the SMEC related to the increased reserve demand. In the real-time market the capacity reservation of the binding constraint is difficult to reflect in future non-binding five minute energy prices. To the extent in RTPD a resource is not CAISO/M&ID/M.Miller/L.Kristov Page 14 April 5, 2011

15 awarded another reserve such as spinning reserve to allow the resource to be available to meet the forecasted ramping requirements, there may be an opportunity cost for such a resource. On the other if the ramping requirement is being provided by resource capacity that is otherwise not economic for energy or other reserves, the opportunity cost would be zero. Therefore it may be appropriate to consider a compensation method to recognize at least the opportunity affects of the additional ramping constraint. The appropriate pricing mechanism for this constraint will be assessed based upon how regularly the constraint is binding and the extent to which the constraint results in lost opportunity for resources. As an interim step, ISO proposes to share information including the quantity of the requirements and the associated shadow price of the binding to assess the appropriate compensation method. 2.4 Allocation of Integration Costs The integration of substantial amounts of VER capacity into the supply fleet poses several challenges that will tend to increase the costs of meeting load and reserve requirements and maintaining reliable operation. These increased operational costs will result from a combination of increased reserve procurement, greater operational demands on conventional resources, increased need for resources that can offer specific performance capabilities, potential for increased uplift amounts, and the expected need to provide some form of capacity payments to conventional resources to supplement the reduction of spot market revenues as spot prices decline due to the lower marginal costs of VER. A central question for this initiative, and the topic of this section, is whether these integration costs should be allocated directly to the VER that may be viewed as causing the increased costs, and if so, what cost allocation principles should apply and what methodologies should be used to determine each resource s appropriate cost share. An overarching question to keep in mind throughout this topic is how the market rules particularly with regard to allocation of integration costs can be used to provide longterm incentives for developers of VER to design new renewable resources that are better able to manage their own variability and reduce such impacts on grid operation. In pursuing this initiative the following are some of the questions that need to be addressed. 1. Which specific integration costs should be allocated to VER? 2. How should the relative cost shares charged to demand versus charged directly to VER be determined? For example, if VER should be charged for a portion of the cost of ancillary services, what methodology would be appropriate and fair for measuring the incremental impact of VER on this cost? 3. Should cost allocation be based simply on resource categories (e.g., technology, PMax), or should it be based on measured performance of each resource during the hours the costs are incurred? 4. If measured performance is the basis for cost allocation, should these costs then be allocated to all resources, regardless of resource type, that exhibit the same performance to a lesser degree? 5. Should allocation of integration costs be limited to the operational and spot market areas, or should we also consider allocating some of the integration costs through the generation interconnection procedures (GIP)? The list above is not intended to be exhaustive; stakeholders are encouraged to help refine the key questions, identify others that have been omitted from this list, and suggest guiding principles that will be helpful in this effort. CAISO/M&ID/M.Miller/L.Kristov Page 15 April 5, 2011

16 2.4.1 Integration costs for VER Imports Under new dynamic transfer provisions the ISO is now completing for submission to FERC within the next few months, VER outside the ISO will be able export their variability from their host BAA into the ISO using either a dynamic scheduling or a pseudo-tie arrangement. The current dynamic transfer initiative did not address the question of allocation of integration costs associated with these VER imports, however, but deferred that question to the present initiative. The ISO expects that the questions and issues identified in section above will be relevant for dynamic VER imports as well as for internal VER. The ISO requests input from stakeholders as to whether there are specific factors or attributes associated with dynamic VER imports that make them different to internal VER and require different treatment or special consideration with regard to allocation of integration costs. 2.5 Modifications to Intra-Day Market Settlements Full Hour Ahead Market A possible market design enhancement being considered to better accommodate the scheduling of intermittent generation resources is the introduction of hourly intra-day markets, cleared and settled for each hour against day-ahead schedules, with actual real-time load and generation settled against these intra-day hourly schedules. This design differs from the current HASP, which schedules imports and exports against the load forecast and commits quick start resources, in that the intra-day market would determine hourly prices and schedules against which all real-time load and generation deviations would settle. Creating such a market with full settlements would essentially change the ISO market structure from a two-settlement to a three-settlement market. Clearing and settling such a market implies the participation of both buyers and sellers; thus unlike today s HASP a full hour-ahead market would be designed to accept bids from internal load rather than simply clear supply offers against a load forecast. An important design feature for such intra-day markets is to specify the identity of the buyers and sellers in these markets, i.e., whether there should be any limitations on participation, which will impact the nature of the potential benefits from implementing these markets. For example, would the design accommodate all three possible convergence bidding combinations: day-ahead to hour-ahead, hour-ahead to real-time, and day-ahead to real-time? Before we go too far in trying to specify potential design alternatives for an hour-ahead full settlement market, it is important to identify the potential benefits that could be achieved through such a market, particularly with regard to renewable integration which is overarching purpose of this entire initiative. As we have discussed with stakeholders in Phase 1 of this initiative in the context of revising the PIRP, the ISO does recognize that many VER would like an opportunity to establish a schedule as the basis for measuring real-time deviations that can be specified as close as possible to the real-time operating hour. It is not clear, however, whether an hourahead settlement market would be an effective vehicle for this i.e., it would probably have to utilize the same time frame as today s HASP, with bids submitted by T-75 and market results posted by about T-60. Moreover, it s also not clear that instituting a full-blown third settlement in the market structure is necessary to accomplish the limited objective of providing a scheduling opportunity for VER. Thus, at this point in the present initiative, the ISO does not see how the cost and complexity of designing and implementing a full three-settlement market structure could be justified in terms of the benefits it would provide. The ISO requests that stakeholders CAISO/M&ID/M.Miller/L.Kristov Page 16 April 5, 2011